Nvidia Is a Must-Buy. Or Is It?

In 2002, after the dot-com bubble burst and Sun Microsystems swooned, the company’s co-founder Scott McNealy highlighted the folly of Wall Street analysts who favored one particular financial metric to gauge a stock’s worth: its price relative to the company’s sales.

Mr. McNealy was musing about the “price to sales” ratio — an important measure of a company’s value relative to how much cash it generates. A high ratio can be justified if investors think a company has room to grow; a low ratio typically signals that investors think the company is accurately valued.

Using that metric, analysts had gambled that Sun’s stock was undervalued even when it was trading at more than 10 times its revenue — a value its business couldn’t ultimately sustain. Even if Sun passed on every dollar it was making at the time to investors, it would have taken them a decade to recover their investment.

“Do you realize how ridiculous those basic assumptions are?” Mr. McNealy told Businessweek. “You don’t need any transparency. You don’t need any footnotes. What were you thinking?”

The current stock market is evoking similar sentiment among some investors, led by the giant chipmaker Nvidia, the poster child of the exuberance around artificial intelligence. On Wednesday, Nvidia’s stock price closed at 27 times its sales.

Nvidia is very different from the hundreds of revenue-rich but profitless companies that the market cheered on in the late 1990s. The company, in Santa Clara, Calif., is wildly profitable: In the final three months of 2023, it generated over $22 billion in revenue, up 22 percent from the quarter before and more than 250 percent higher than a year earlier.

But does Nvidia have enough room to grow to justify such a steep price-to-sales number, or is it magical thinking on the part of over-excited investors? Experts are divided.

The high price-to-sales ratio is rooted in a firm belief among many Nvidia enthusiasts that the company will keep growing because of its critical role in artificial intelligence. Even if a ratio of 27 times sales puts a big growth expectation on the company, many investors still consider Nvidia undervalued because they expect it to keep generating more and more cash — until eventually, the price-to-sales ratio shrinks to the level of a more staid, corporate behemoth.

That has already started to happen. Before it reported fresh earnings on Wednesday, the company traded with a ratio closer to 30 times its sales. In June, it was above 45.

“The numbers have got so big, so quickly,” said Stacy Rasgon, an analyst at AB Bernstein who covers Nvidia. Mr. Rasgon still expects Nvidia’s value to be “materially higher” in five to 10 years.

But Nvidia isn’t the only company causing consternation, even if it is the most eye-popping. Microsoft, Advanced Micro Devices and Broadcom are among the companies that have seen their price rise above 10 times sales in the past year, as beneficiaries of the general excitement around A.I.

For some investors, uncertainty over whether the bet will pan out makes the high price of stocks like Nvidia off-putting, especially when there’s a lack of clarity around the path of inflation and interest rates, as well as political uncertainty from Ukraine, China, the Middle East and at home ahead of the presidential election.

“What return are you actually getting for taking all that risk?” said Matt Smith, an investment director at Ruffer, a London-based fund manager.

Another popular metric, the price-to-earnings ratio, shows the S&P 500 is now trading at close to 23 times the collective earnings of the companies in the index. Excluding the pandemonium around the pandemic, the last time the ratio was that high was just before the market stalled in 2018. Before that it was when the dot-com bubble burst.

For stock prices to keep rising from here, either profits have to keep growing or these metrics favored by stocks pickers would have to push even further above their historical norms.

“Valuations are already historically rich,” said Jordon Brooks, co-head of the macro strategies group at the trading firm AQR. “And we would be talking about them expanding dramatically from here.”

However, relying on snapshot metrics oversimplifies whether or not a stock is still good value for the money, said Aswath Damodaran, a professor of finance at the Stern School of Business at New York University, where he teaches about equity valuation.

Amazon in January 1999 was trading at a stock price that was more than 40 times its sales. Since then, its stock price has risen an average of 15 percent annually. Its revenues have grown even faster. Today, its stock is just three times its sales, and it has been one of the best investments in the S&P 500 over the past 20 years.

Nvidia could be the next Amazon and fulfill investors’ growth expectations. Or it could end up more like the dozens of computing companies that came to prominence in the 1980s but didn’t last till the new millennium.

In 1982, Commodore International sold the second-most-popular personal computer — the Commodore 64. By the beginning of 1985 it had lost its competitive edge and its stock price had slumped from over $100 to less than $20. Less than a decade later the company went bankrupt.

“People said PCs were going to take over the world,” Mr. Damodaran said. “They were right. But what they were wrong on was all the companies that were making PCs in the 1980s that didn’t make it.”

The same is likely to be true for many of the companies swept up in the A.I. boom, he added.

Similarly, when it comes to broad indexes like the S&P 500, simple metrics don’t tell the whole story. Remove the so-called Magnificent Seven stocks, like Nvidia, whose size has had a big impact on the overall performance of the S&P 500, and the index looks much more modestly priced compared with its past performance.

Picking the Amazons and dodging the Commodores still isn’t easy.

Such analysis inherently relies on assumptions about the future — a company’s future profitability, its future competitors and even the future of the world it will exist within. That uncertainty helps explain the wide range of expectations among Wall Street analysts, with the most pessimistic seeing Nvidia’s true stock value to be closer to $400, not the $674 closing price on Wednesday, while others think it should trade over $1,000.

Mr. Damodaran considers such high expectations “unrealistic.”

“It is the nature of the beast,” he said. “We think we can do more than we can. When a big change is coming we overestimate.”